Vous êtes sur la page 1sur 19

ORIGINAL

l\lethodological

Preconditions and

ProblelTIs of A General
of Accounting

HE last
fifteen years of accounting
history seem to be part of a transition period replacing the loo:oe traditional approach
hy mun: rigufous methods .. \lthough spanning oyer consideralJle
time this transition continues, most likely
transgressing
into the eighties, and promises to cause a major break in the evolution
of our discipline. A change of spirit, much
more than computerized
technifjues, is the
main force behind this trend. The distinction between
'\'hat I call the modern,
rigorous or generalized approach,
on one
-;Ia~<-;~il(lthe loose traditinnal
or particlIlarizeda[>proach:
on"the ;)-tll;;:;5 st~i:;'1a~,[Tzed--by-tilCtonowing
five points:

'~1.

Formulation
and utilization
oi well
terms and empirically llleaningful
concepts versus employment
of vaguely
II tkscribeu
cxpressions :lnd nUllllpcraLiona[
~oncepts.
'\,1'1'''' 2. Adapt:J.tion of general scien tilic tools
'1 ",':d .. ,and methods
irom mathematics,
philoso'JJl4~" "~'.,,1' phy, economics and the behavioral
sci;1 \
ences to accounting
theory versus par\ticul.ar~za tion and u tiliza tion o.f a n:l rrowly
speCialized conceptual
framework.
\
3. Orientation
toward specific accounting and management
information
models
y
for specific objectives
versus dogn;alic
acceptance
of a unique, overall or undefined purpose.
c

" <,.,-' defined

~J

Theory

.J.. Systematic testillg procedures through


which alternative
accoullting mouels and
hypotheses
(for one and the same objectin) elll be tested as \0 their relevance,
reliability, accuracy, cnicicncy, timeliness
or perhaps
overall prolitahilityl
versus
men: testing of stewardship and fulfillment
of convcntions.
...-- .... S. Integration of ,;pecii~l:accounting areas
to a coheren t en t ity \TrSUS collecting
loosely connected
conwntions,
dogmas,
The author wishes to npn'"
hi, gralittllie to the
("allad:t Coullcil as wel! :<s the l.''';\Trsity of l.lritisu
Columbia (Killam Swinr :\ward) for supporting- this
project.
I In the s'''mner
or [')(,7 tl . ,Luth"r made a public
plea at the Second Internalio",,1 Cunkrence on .\ccnuntinc: Education (see f>r"cmfig~s,
PI'. 1653) that
o'uch systematic testin!!: pnJCed",,'s and their epistemolo~ica! rnund:<tinns be exp!"red (i"<lt"ole:1S or the present paper), This plea seem,; tu h:<\T r<lllllUresponse in a
series of pertinent artide:;; Thomas If. \Villiams and
C. JI. Gril1in, "On the ~ature of Empirical Verification
in :\ccountin;;," .lbaClts (Decemher 1')69), PI'. 1-13-i8;
Robert R. Sterlin.e:, "On Theory Conslruction and Veri!ic,,-tion," 'I'm: !\CCOUl'T'~G Rnll':w \July 1970), PI'.
4l-l-57; T. II. Williams, W. II. Heaver, D. L. ~rcnon:dd, K. F. Sknuscn an,/ R. ~ierlill~, "Report of the
t ~onllnitlt.:e on Accoun lill~ 'rhenry Conslruction and
Ycriticat;oll," THE .\ccnl'''Tl~G RE\'\EW'-SUPP!Clllent
l" Vol. 1(1,1971, PI'. 50 7').

Richard JIattessich is Professor, Faculty


of Commerce aud J)llsillCSS .Adlllinistrati01~
at The University of nritisil Columbia.

rllks ;1111\ iSl,btt-d jlarticlIbri/,<'<\ Ill'HkL,:!


Frum time to time the nl'l'<1 arises to
l<l\,k ;d P;\st rl'sl'arch acti\'it v from the
]>ir<1':-;pn:-;pcctive with the ain; of an ov"rall S\ln'v}', :)u(h is the intt'nt of th\' 1<)1I,y\\'in~ investig~ltion which tries to lit recent rC5'earch efforts into this live-point
pattern, ;ll1d \\'hich examines major melho<IIJ]"gi(-:t! and rdated prohkl11s slill 10 be
soh'cd lll'lore lhis phase oi transition
is
('(jl11pktl'd, l'rior to such an investigati(lll
it lll;lY lie oppqrtune
to av()id some misu ndns ta nl Iings.
It is neither my intention
to deride
trtlJitiOIl<1l
tlcco/llitillg
,\'hich I recognize as
t he basis of [u rther encka \'nr5, nor do I
I'll-ad f"r eliminating
va IU!.: judgemcllls
f rum our \liscipline. IIowever, till: persisknl'\~ of value judgements
raises tl1e::<:rious question "'hether modern accounting
\\'illl)(_, as dogmatic as traditiun;t1 accounting is .. \ normative approach which npliC:ltl'S ib value juJgem<:nls
alld r<:duc<:s
th<:m to more basic levels, which furthermore applies systematic testing procedures
to reject inadequate
systems, could well
be regarded as a nondogmatic
approach.
It would be extravagant
to regard accounting as a pure or cognitive science.
But neither medicine nor meteorology nor
epgineering are pure sciences; they are ;l0t
concerned with finding scientific laws out
with serving specific practical purposes in
a pragmatic fashion. These disciplines are
generally recognized as a p pl ied scieuces oecause they operate with such scientific
concepts as hypotheses, diagnoses, models,
theories
and systems,
and endeavor: to
subject these hypotheses,
diagnoses, models, theories and systems
to systematic
testing procedures with the serious intention to eliminate those theories which the
accepted
criteria indicate
as inadeq~ate
for their practical objectives?
Since progressive accountants
and manucrement
ino
formation
systems
experts
also operate
with models, theories, systems, and the

lik,- ;11\,[ illl'- Ih,'." Ifill :I:-,-pin'!o\':;llclnll,r,'


rigorou:-,- It-sling pron'dures,
tlw g:lp iw1\\'I'l'1l ;1\ \\J\llltill,~ and otll<:r ;1]ll'lil"d'\ 1\'I\I'\:S
nl:l:' 1101 Ii\" as great as S(JlIIl: of I,ur
~-\d\l.':lgut".-:'

~lI,~~gt~l.

The ultimate purpOSI: of accoullting


i;;
to provi(k managcnalllllormatlOlI
syskl1ls
~;;Itis[a\'tory or "'Tn optimal
[or S{WIilic

~Th;;tlTii:;

shlluld

hi:

done

Ihn>1lgh

syslematic
ksting
and by means or all
inslrunll'lllal
theory
COIl\\lH:nSllr;t!e 10
ueeds as wdl as scientific meallS al Illlr
disposal, hardl)' appears to h<: an unreasonable quest.

(;).1.

TER~lli':I)\.I)(;IL\L ;\,,(J

CO:--:CI':I'TU,\L

!)IFFICULTIES

f)t/iJIcd

I'aslls

Uwlcjilled

Terms

Traditional
~\l'coullting, as must uLhn
acadl'mi(' disciplines, partly borrows ,'011I'l:pls i rlJlll lIei.~hburing (lisciplillcs
;tllIl
l':lrtly <T(;tlCS its OWll \'OllClVtn:l! aj'\':Iratus. Yet, a(:!'ounting thl:ordicians
rar<:ly
make all drortlo
indicale dearly whidl of
their terms are primitive (hence borrowt.:d
from outside) and which are derived from
these primitives
by means of nominal
delinitiolls
within the accounting
theory
proper. This vagueness is one reason why
the boundaries
of traditional
accounting
theory and its subareas are blurred and
2 A strict distinction
between a "specific model" mentioned under (3) and an "isolated particularized model"
mentioned under (5) ou~ht to be made. \Vherea5 the
former is cmbedded in a theoretical framework and thus
has common bonds with other specific models, lhe particularizcd model lacks lhis connection and is theoreti
cally isolatcd.
> "As we shall see in some detail, applied research has
the advanta!;c of bein!; able to formulate criteria of its
own efftciency in terms of the objectives for which the
problem is being investi!;ated. Because of its lack of
specific objective5, pure research cannot formulate such
criteria as explicitly," Russell L. Ackkott, SciwtiJic
ill el/lVd-o/Jtimi=ill~
A ppliai Researc!/ [)ecisions (John
Wiley & Sons, Inc., 1962), p. 24. Also "In scicnce,
whether pure or applied, a theory is both the culmination of a research cycle and a !;uide to further research.
In applied science theories are, in addition, the basis of
systems of rules prescribing the course of optimal practical action," Mario Bun!;e, Scientific Research II -Th4
Search (or Truth (Springer Verlag Co., 1967), p. 121.

-f1)72

orc
he-

-(1-

\-

1\

-, is
ms
'<hie
!gh
yan
to
-'Jur
ea-

1\

\1

"
.!

I
I
I

I
Ii
1l'r
;Ilnei
Iy'

"f
",d
." ,rn
l~tl
,ry
hy
ng

nd
tD-

:1,'1"

l;lr-

-rtie
"'~13.S

its
!he
.,f
",:ldl

(C,

n:l

.:h.

:"f
;lC-

rh~

I.

why accounting itself is usually described


instead of defined. For example the well
known "delinition"
of accounting
ofiered
by thc Committee on Terminology
of the
American
Institute
of Certified
Public
Accountants~ is neither a nominal, nor an
analytical,
nor an operational
definition;
it does not set sharp boundaries
to the
term "accounting"
nor docs it prescribe
operational
niles by means of which one
can test whether something is an accounting system or not. Indeed, the "definition"
would fit other areas of business administration equally well.
Although
the failure
to distinguish
between
primitive
(undefined)
and defined terms leads to confusion, it might be
a lesser evil than the failure to discertl hetween interpreted
and unintcrprcled
concepts and calculi.
Illter preted

1'CTSIIS Uni I/ta prdt:d

Co 1/((' pts

\Vhen-as a ddinilion
is a n!:lIi"nship.
between tt.TlllS or I,d wn:n a krill alld its
constituents,
an intcrpret;tlion
is a rebtiol1ship uclwt:t:n a phenomenon
(iact or
idea) and the term (symhol) reprcsenting
it. General concepts require specitic anti,
in many cases, empirical intl'rprclations.
In accounting,
uniortunately,
thc dis[inction bclween
inlt:rprcled
and llilintnpreted concepts is even more nl.'glecled
than that between detined and urrdciined
terms_ COllccptS like income, ~:'c(/lth, ~'(I!/le
lllld -measure Ilre treated as tllO/lr)1 they 7:!ere
'well interpreted
COl/ecP/s, hut at !lest an
interpretation
is merely implied, and in
the case of scholarly controversies
the two
opponents
often imply different
interpretations, usually without rcalizing it.
In this regard the example of the prohahility concept offers a valid analog frolll
which accountants
can learn much. For
considerable
time hot controversies
were
fought over the question
whether
probability is an objective-empirical,
an objective-formal,
or a subjective-empirical

concept. This controversy


was resolved
afler Kolmogorotf created the general and
thus uninterprcted
probability
concept
stipulated by his axioms of the probability
calculus. Then differellt Titles of iJlterpretations were' devised in such a way that
under one set of rules the relative frequency
concept
(an objective-empirical
one) emerges, under another set the degree
of conlirmation
(an objective-formal
concept), and under a third the Bayesian
probability
concept (a subjective-empiri~
cal one). Later even more interpretations
were devised.s \Vould it not be possible,
in a similar fashion, to devise certain conditions or common traits for the concepts
of income, wealth, vahll: and other acCOllllling conn:pls in (lrdn to create lirst
gencral uilinterprdni
1"<1Iln:pts and then
rules of interpretations
for specilic needs?
In this way, much clarity l'Ould he hroughl
into our disciplilll' ;lIHI 111;111)' a futile con"
t r(lvlTsy c011ld he ;1v id."d.
1':lldnl,'o/lrs
tioll

:l'o'h"/I,l

fjdta

('ollceptl/llli:a-

Systematic
attempts
to <Teate a series
of clearly delined tll1inlcrpn:kt1
concepts
of the major accounting notious and alternative inkrprctalions
(for specilic or standardized nl'eds) of l';lch cOllcept have not
LJeen made in our discipline. But in an informal, vague and incomplete manner such
interpretatioJls
han; becn customary
in
traditional accounting for a long time. One
could wl'l! argue Ih;tt the acquisition cost
basis, the market
value basis and the
present value basis are nothing but "interpretations"
of a general
unintcrpretecl
value concept.
Hut the essence of the
Illdhodological
achiev('llH'nl
of distinguishing hetween uninterprdcd
and interpreted concepts and theories (calculi) lies
, ..f ((ollllting Terminology Bullciill No. 1 (American
Institute of Certified Public Accountants,
1961), p. 9.
6 See Mario
Bunge, SciOli/fie Research I-The Search
for System (SprinF;cr Verla!; Co., 1967), p. 428.

in the spccifietlioll of the conditions rllaracterizing every unin terpretcd concept.


These rules of interpretation are rarely, if
ever, spelled out in accounting. Thus the
practice of distinguishing various species
of one anu the same super-concept as encountered in our discipline, is merely a
first step in the direction of fulJiliing this
important methodological prerequisite.
However, credit has to be given to
several academic accountants for having
concerned themselves with this or the
closely related problem of oper;ltionalism.
Bedford, recommemls the utilization of all
operational income concept:
The fact is that there is a need for one overall general concept of income in our society. The formulation of the set of operations to be used in developing such a general concept of income is by far
the most difficult aspect of such ;t process of
theory formation.6

Yet, he (apart from distinguishing between business income and "income as a


generalized means to gratify a variety of
human needs"i which he deems beyond
the concern of accounting) seems to regard
the generalized (business) income concept
as an interpreted instead of uninterpreted
concept, and thus docs not aim towards
further interpretations. For him the operational rules are what we called the cOllditions of the general concept, while in modern methodology the operational rules
would be equivalent to the rules of interpretation of the specific concepts.3
Devine, too, recommends in his essay
"Principles, Theories, Systems-:\gai n"
that "principles be operationally Jc[liled
and treated as guidelines."~ Explicit mentioning of "semantical rules" (i.e., rules of
interpretation)
in the recent accounting
literature is made by Sterling in his paper
"On Theory Construction and Veriiication"IO and in the" Report of the Committee on Accounting Theory Construction
and Verification."Il Yet, the actual significance of the rules of interpretations for

cn:atillJ: su!lconn'ph and slIhtlworits frolll


the general concept or theory respectively,
is not emphasized in these writings.
To attain better conceptualization in
accounting and management information
systems, comprehensive and systematic
taxonomic research is indispensable.l~ Such
research will have to formU!;lte oath th~
conditions of each of the uninterpreted
accounting concepts as well as the rules of
interpretations of all the specific and standardized subconcepts. Such classi6.cational
rese:trch is tedious ;md hy far not as glamourous as model building, allli it might
take some time before the Linnaeus of accounting emerges. Ho'.yever, if biology was
in a position to classify over one-and-ahalf million species of plants and animals
within a complex taxo!\omic system of
kingdoms, phyla, classes (wi t h subclasses),
orders (with suborders and sections),
families, genera and species, then it should
not be impossible for accounting to produce a dassificational structure of a few
dozens or hundreds of concepts and subconcepts.
~\I)Al'TATION

JIathcmatical

OF SCIEKTIFIC

and Ecollomic

METHODS

COlltributions

The greatest progres:; of the transition


period, maJ~ thus far, lies in adapling economic ideas and ma.thematical methods or
'Norton
Bedford, Illcome
/)ctul1li"atioll
Theory
(A{ldison-\Vesley, 1965), pp. 8-9.
, Ibid., p. 11.
S Bun!;c, Scientific
Research i-Tire Scarch jur System,
Pl'. \.)<)-53.
\J Cart'r.
!)cvinc, HEssays in A{'cu\inling,H VoL lIlt
Tallahassee, 1971, p. I. (i\lil1le"l~ra!,hnL)
10 Rohert
Ie Stcrlin!;, "On Theory COllotruction and
Veriflc:tlion," 1'111: Accu1J:sn:<G REVIF:W (July 1970).
Pl'. -+44-57.
II Williams et ;tl., pp. 50-19.
12 John
T. \Vheeler, "Accounting Theory ;tnd Resc;trch in Perspective,"
Tin: AccilUt>Tl:SG
REVIEW
(Janu<try 1970), who :tIs<>pleads for "more effective
classilicaLion schemes" of accounting (I'. S). See also
Oracc Johnson "Tow;trd an 'Evcnts' Theory or I\Ccounlin!::' Tm: ,\("cou:-:n:-;c; j,EVIF:W (Oct"her 1970),
who a5scrts lh~t t~a prirnary need frlr a.n accountingtheory is taxonomy ... " (1'. (14), and idem. "On
Taxonomy and Accounting Re:iearch," TIIF; ACCOUNTI:SG REVIEW (January
1972), Pl'. 6+-H.

techniqu
of these
known, J

and sum
a metho
ingly we

foHowiof
investm(
studies t
network~
.,: tion and
statistica

. informat
,j

.i
I

Prcsen
. ,

tions are
:ipplicati
tional ac
rived fro:
merit

to

by meal
fundamel

interpret
broader:;

words i~1
.'
..
;"11
lffiportan
theory e:;f
radiC3l,'i
stil(itI~
quenct$.

h}"C~
lir.g. an.ii
. .Bienn:m
~'
r.:-tz , .~
Ct'n~~

pfO.l.dt:,

Thc!4t
of~
&rtidef;
a '"0..u 'It:

~,e .,_:;~
u;t"Il~

('<~

a:~~f!!

U:~~

es from
ctively,
tion in
-mation
tematic
:.12 Such
oththe
~rpreted
rules of
I1dstancational
as glamt might
us of'aclopz ,vas,

,e-, l-aanimals
'stem of
)classes) ,
;cctions),
it should
to proof a few

and subETllODS
'Iribl/tiolls

transition
pting ecolL,thods or

techniques to accouilting. Since the details


of these achievements are sufficiently well
known, I can restrict myself to classifying
and sumnnrizing these achievements from
a mcthodological point of view. l\ccordingly we might distinguish between the
following categories: (1) Prcsent value and
investment
calculations, (2) structural
studies by means of matrices, vectors and
netw'orks, (3) linearvrogramming, simulationand computer,applications,
and (4)
i statistical
applications and valuation of
,inwnnationY
(.
(-'j Present value and investment calculaStions.are
no novelty in accounting. Their
applications have to be credited to traditional accounting but were ultimately derived from economics. It is Irving FisiH;r's
merit to have interpreted capital theory
by means of accounting concepts. The
. fundamen tal significance of this Fisherian
interpretation lies in a new and much
broader vision of our discipline; in other
words it was the first and perhaps most
important step to\vard a. more ~eneral
theory of accounting. But it was such a
radical innovation that some accountants
still refuse to grasp several of its consequences. N everthcless, strides were llla(le
by Canning,!lb.ter by Moonitz and Stachling, and in more recent times by :\Ibach,
Bierman, Hansen, Honko, Jordan, :Moonitz, Sprouse and others to c\abora. te the
consequences of the present value approach within the conlines of accounting.
The justification to include here this kitHI
of development is best supported by two
articles that appeared almost a uecade
i ago.IS Both Corbin and Philips spoke of a
"revolution" in our discipline; indeed they
concentrated chiefly on the economic valuation facet just mentioned, without enlcring into other revolutionary aspects then
visible on the horizon.
The insight that the matrix notion is a
general and convenient means for depicting the structure of the basic accounting
I

".truction 'and
-Uuly 19iO),

'Ory

and' Re-

ilY;'
i

REytEW

effcctive
,./. Scc 'also
ilcory of Acetober 19iO),
~Jl accounting
't idem. "On
'HE ACCO~T-

models had far-reaching consequences for


business accounting, as well as beyond it.
Above all, a rectangular matrix reveals the
double c\assificational structure of accounting in a more generally understandable fashion than does the language of
traditional bookkeeping; and second, matrix algebra lends itself to the solution of
many allocation problems and related
issues of micro- and macro-accounting.I5
Some experts regard this matrix mode a
ncw "paradigm" by means of which accountants and economists nowadays visualize the basic structure of economic
flows. This claim may be reinforced by the
"official" adoption of the matrix mode in
national incolllcaccounting.
The United
0:'ations, for example, hased their recent
publication A System of National Accounts17 entirely on the matrix mode of
accounting .
A further advantage of this mode grows
out of its mathematical generality. Due to
the latter the matrix mode can easily be
con verted in to the vector mode, or network mode, or any other equivalent (yet,
occasionally more convenient) way of
looking atan input-output structure. Since
most accounting and management information systems possess such an inputoutput structure this new paradigm may
13 For n. categori.zation
of traditional research see
Norton M. Bcdford, "Rcscarch Projects in Accountin~," in Tnpics in tl cr"'lIllill~ and Plill",iH~, Cll. hy R.
:\lattessich (Mono:,:raph "'n. 5 of the Faculty of Commcrce and Businc~s :\<ll1linisl rali,m, University
of
British Columuia, 19i I), i'p. 1&-25.
.
" J. n. Cannin:.:, The l:collumics of Ii ccO/wtallcy
(Ronald Prcss, 1'J2')).
" Donald A. Corhin, "The Revolution in Accountinp;," THr; i\CCOF:'-:TI:-;C REVIEW (October 1962), pp.
62&-35; and G. Edward Philips, "The Rcvolution In
Accountinl';
Theory,"
TilE
A.CCOUNTll'G
REVIEW
(Octobcr 19(3), Pl'. 6c)6-iO~.
" R. Mattcssich. "Tm\'anls a Gencral and Axiomatic
Foundation of :\c,otmt:U1C\' -\\'ith an Introduction to
thc Matrix Formulation of Accounting Systcms," Accoulltillg Research lOctober 195i), pp. 328-55, offering
a gcneralized scheme applicable to micro- as well as
macro-economics.
t7 Srudies
in J[ cl!lOds, Series F, No.3 (Oert. of Economic and Social i\fTairs, SLltistical Ol1icc of the United
Nations, 1968).

_. ____

. _T ______
i

i:

i
become an indispensable element in the
design of a general theory bf accounting,
even where no actual double classification
is practiced.
. f
Apart from these methodological aspects
which have been treated extensively clsewhere,18 the more algebraic aspects of this
new paradigm have inspired a series of
exciting accounting contributions on mieroeconomic input-output. models, their
theory and further generaliiation. Starting
from the interindustry model and the insight that Leontief's basic idea of t'llterdependent commodity
transfers between
production areas- need not be restricted to
!he macro-economy, several authors made
significant contributions by applying matrix algebra to traditional cost accounting,
budgeting and other areas.19
Linear programming, simulation and
computer applications in several respects
are related to the research area indicated
in footnote 19. They, too, constitute a
comprehensive body of mathematical studies that greatly contributed to more rigorous analytical thinking in accounting.
Both linear programming and the matrix
calculus are part of linear algebra, and
although the matrix application of accounting is strongly oriented toward allocation procedures, it too may be geared to
planning and budgeting, like linear programming and systems simulation.20 Thus
a peculiar feature of our "transition
period" is the milch stronger orientation of
accounting tOLI.!ardprojectio ilS of future c(:onomic et'ellts. Traditional accounting did
not lack this element (already iifty years
ago did 1\'leKinsey lay the foundations to
a systematic and comprehensive business
budgeting),21 but during these live decades
budgeting has never been more than a side
issue of minor importance. Even today
some experts wish to exclude budgeting
from accounting proper. But it is this neg~
leet of budgeting which has deprived traditional account1l1g of its ~nost important

"~-I

-.

task, namely to answer the question "How


well could management have performed if
they would have allocated resources in the
best feasible ''lay?'' Numerous recent developments of linear programming and
system simulation, however, indicate that
the control function as well as the exploration of alternative factor combinations, by
means of budgeting and management information systems, is becoming a dominant issue of modern accounting.22 Budget18 Richard
Matlessich,
AccOtmti,1g alld AllaIytical
Methods (Richard D. Irwin, Inc., 1964) and Die wissellschaftlichen Grmldlagen des RU!l1lWlgswesens (Bertclsmann University Press, 1970).
Neil Churchill, "Linear Algebra and Cost Allocation: Some Examples,"
TilE ACCOUNTING REVrEW
(Octoher 19M), PI'. 89+-904; Thomas 1L Williams and
Charles 11. Cril1in, The !of athematiwl lJimwsion of .ticCO,,,,tinl: (South Western l'uhlishin~ Co., 19M} and
"Matrix Theory and Cost Allocation," TilE ACCOUNTING REVIEW (July 1964), Pl'. 671-i8; Rene P. Manes,
"Comment
on r.latrix Theory and Cost Allocation,"
TIrE ACCOUNTING REVIEW (July 1965), Pl'. ~3;
Shawki M. Farag, InfJ1ltOlltf!ftt Allalysis: Application
to Business AccOttnting (University
of Illinois, 1967)
and "A Planning :Model for the Divisionalized Enterprisc," TilE ACCOUNTING !{r:VIEW (Afril 1968), PI'.
312-20; Yuji Ijiri, "An Applicalion
0 Input-Output
Analysis to Some Problems in Cost Accountin"," .llanagemmt Accounting (April 1968), Pl'. 49-61; John Leslie
I.i\;ngstone,
"?\1atrix Algebra and Cost Allocation,"
TIlE AccomlTl:SG REVIEW (July 1968), Pl'. 503-508
and "Input-Output
Analysis
for Cost Accounting,
Planning
and Control,"
TIlE ACcom,;TING REVIEW
(January 1969), PI'. 48-64; Gerald A. Feltham "Some
Quantitative
Approaches to Plannin~ for Multiprociuction Systems,"
TIlE ACCOGNTING I{EVrEW (January
19iO), Pl'. 11-26; Trevor E. Gambling, "A Technological Model for Use in InputOutput
Analysis and
Cost Accounting,"
J[ allagement Aceollllti"g (December
1968), pp_ 33-38; Trevor E. Gambling and Ahmed
Nour, "A Note on Input-Output
Analysis:. Its Uses in
Macro-Economics
and Micro-Economics,"
THE AcCOUNTING REVIEW (January
1(70), Pl'. 98-102; B.
Sigloch, "Input-Output
Analysis and the Cost Model:
A Comment,"
THE ACCOUNTINGREVIEW (April 1971),
pp. 374-79; John Butterworth
and Berndt ..Sigler..h,
"A Generalized }''1ulti-Stage Input-Output
Model and
Some Derived Equivalent
Systems," THE ACCOUNTL'lG
REVIEW (October 1971), Pl'. i~16.
2. On the other hand many linear programming applications in accounting are related to allocation problems
via the broad issue of transfer pricing. Beyond that,
there exists an interesting attempt to apply linear proRramming to overhead allocations: see Rohert S. Kaplan
and Gerald L. Thomson,
"Overhead
Allocation
via
:Mathematical
Programming
Models." THE ACCOUNTING REVIEW (April 19i1), pp. 352-64.
n James O. McKinsey,
Budgetary Control (Ronald
Press, 1922).
22 G. G. Mueller,
ed., A Nr:w Introduction 10AUOfInIing-A Report of the Study Group Sponsored by the
Price Waterhouse
Foundation
(\971) recommends
to

MattI

Cff

ing ar
made
to prt
Green
that I
tive a
aecou
rect,

they
,!

abili~
predi<
assun:

other
Greet
that I
pothe
hypo!
plied
ment4
whicl
testin
this Ii
Sta
of in.'
enh::.
rese;
trUJ.l1

distiL

sufis

(esPO
ing);

bt.hz,
lioOt

tian~

..

.. ,:l,~;

:cocml

'i !:it!

'OtW;

'How
lcd if
n the
deand
~that
)loralS, by
1t

in-

c1omildget',llyticaJ
)ie uoiss (ller:\llllca~J-:VIE\V
.IIlS

'''d

'I 0,

>-if and
'l'Ot;~"T-

:-lanes,
!";ltinn:'
(,\()-.1.\;

,'/'",1;.11,
. 1%7)
Icntcr, pp.
, 'utput
," Ma1l-

l.t~l\r
,'a lion."
';0.1-508

III

!,untin~J
1,EVU':\V

"Some
,;>roduc!

;:ltlU"rr

'rechno"is and
t'ccrnher
,\hmccl
lOses in
'Ill-:
Acl\ll; B.
:-Iodcl:
;) l~~t),
~i
h,
,del and
,!(':-;Tt~G

"g appli,rnblems
:HI that,
,e"r pro- '
!'apl~n
VIa
.CLva~"T-

'lCCCHlnl-

1 uy the .
,ends to

li"liiiiiil~IW5liiij':'

ing and similar projections into the future,


made in the applied sciences, correspond
to predictions
in the pure scicnces. Thus
Greenball and other authors
who claim
that "it is incorrect to speak of the predictive ability of an accounting method or an
accounting number"23 might well. be correct, but they say only half the story if
they neglect to point out the project~ve
ability (i.e., the capacity
of a fictitious
prediction
under more or less unrealistic
assumptions)
of accounting
systems and
other systems
of the applied
sciences.
Grcenballmay
abo he corn:cl byassertinJ..;
that an acclllllltin~ mcthlld is not a "hypothesis" as long as he means a cognitivL:
hypothesis. But accounting
and other applicd sciences are crowueJ
with instrumen tal or purpose-orien ted hypot!H'S('S21
which arc llO less slllljl'd
to SYSklll:llil'
ll's(in~ pron'dlln's
(sl'e Ihe 1;ls( sed iOll of
this paper).
Statistical
applications
and valuation
of information
is the fourth area ~rcally
enhancing I'igur and qu:!lity of accounting
research. This group spans a broad spectrum of which several subareas might be
distinguished:
First, the application
of
statistical sampling and hypothesis testing
(especially in auditing and control charting) ; secOIH1, the econometric study of cost
bchaviour;z.
third, deci~ion and information theory anti the evaluation of iniormation;2~ and linally, all other statistical
accounting applications
difftcult to classify
otherwiseY
Statistical sampling and con trol chart ing
are subareas of m'ost immediate
practical
usefulness, while the econometric determination of cost curves and the like seems to
be the most neglected one within accounting. The third subarea, however, the evaluation of accounting
information
touches
the very core of our discipline and appears
to have the greatest future potential
(for
further details see a later subsection "The
Value of Information").

iii?i-lilim.iiE7iiMliIiiil1i-iii-ii1--

i1-'IIIiZiilril- ~iiiilliii.tllli!-iiiii'lin.I"- ttirnilll

111easurement

Theory

finally, the application of basic concepts


of modern measurement
theory ,to ac~
counting may contribute
substantially
to
conceptual clarification of several accounting issues. Accounting describes past and
occasionally future events, and thus is a_
special kind of measurement
activity, at
least for those who identify measurement
with quantitative
description.
The conceptualization
of measurement
theory goes
back to 1\1. von Hclmholz and Norman
Campbell, hut assumes special significance
in the social scil'nn:s with the conlrilJutions
of Stevens and other social scientists or
mathematicians.~~
reserve

considerahle

"pace f"r

HlId.l~t1 as. a ('olnpll-lwIIStn'

5:-\. 'rltt' "'{qu)rt oi


IlllollH:tlinn

Illl-

SvstelllS,"

1 he

('lllllltliltIT
'rifF

discussion of "The
I a'vin-," pp, S;

I'bhllill!:

('11

;\(,cllllntlU~.~
alld

.\IIfq::-';"'I~(:

{{EVil':\\'

SlIl>l'll"Il:"tto'Vol.
Ih. 1'171. "-'l'licitly !,oints lIut wilh
re":tru to traditional uccountill::: (in contrast to modern
ac~ounting) that "the inform:ttion was not d~signed for
pl:tnnin::: purposes nor nece<s:lnly for measunn::: performanc~ a~aingt or~ani7.:tlionaJ
objcctivC5," p_ 291.
" i\l. N. (;rcenhall"The
('lcdictive-AI,ility
Criterion:
TIs Relevance in Evalu:Ltin~ .\ccountinl; Data," Abacl/s
(June 1'.171),1'_ 6_
. .
14The essence of these ills/rumental
IJ)'po/lteses IS ,!lscussed in detail in Richard :\l:tttessich,
hIS/mlllcl//ill
Rt:IJ,5<Jllil/g alld lJecisillll .')ystel~ls-A_ Jf e/IJodology o/litc
Admil/;strcttive i/lld otlter A pplted SCiences. forthcomml!;.
~5 Joel
Dean, J[a'wgcrial 1':coJlomics (Prentice-Hall,
Inc., 1'1.11), and J. Johnston, Statistical Cost Analysis
(:\lcGraw.llill,
Inc., 11)60).
:. See especially G_ A_ Felthanl, "The Value of Informalilln "TilE ACCOUNTING
REVI};W (Octoher 1963), pp.
(>i';'1--96;Baruch Lev, ,I ,'c01I-11/illg
lIllIllll/orma/ion
Theory
American t\cclluntinJ.( '\S50C_, 1')(,1)); G. A. Feltham and
J- S. Demski, "The Use of l\lntle15 in Information Evaluation," TilE ACCOUNT1:-lG REVIEW (October 1970), pp.
68-1--96' N. C. Churchill, G. A. Feltham, T. J. Mock,
T. R. Prince, H. M. Sollenherger, F. Kaufman, "Report
of the CnmmiUce on An:Ol!nlin~ and Information Systems," TilE ACC0UNTlIW RE\'IEW-Supplement
to Vol.
46, 1971, PI'. 289-350, T. J. :'-fock, "Concepts of Information Value and :\ccountin:-:," TilE i\CCOU:ITING lZEVIF.W(Octoher 1970, PI'. 765-78.
71 For example,
Rooert E. Jensen, "A Cluster A~aly.
sis Study of Financial Performance of Selected Busmess
I'irms," TilE ACCOUNTING REVIEW Uanuary 1971), pp.
36-56, and many others .
S. S. Stevens, "On the Theory of Scales of Measurement," Scieltce, Uan.-July
19,16) pp- 677-80 and many
suhsc'luent
articles;
C. H. Coomhs,
"Theory
and
l\letho<ls of Sociall\easuremcflt,"
in Research IIfethods
i1l the Behavioral
Seiwas,
cd. oy L. Festinger and D.
Katz (Dryden Press, 1953); C. H. Coombs, H. Raiffa
and R. M. Thrall, "Some Views on Mathematical
Models and Measurement Theory," in Decision Process,

-----~':':

. ::_.::_.;'I:;,:-:.:;:;

m"lill.~Tl~;:,,;:;~;i~i:
.i~l:

t'f~',1';)1
-31:.;

';!

Stevens'
scales were fIrst applied
to
accounting in 19S9~9 and further concepts
of modern
measurement
theory
(e.g~,
"measurement
by fiat") w~re introduced
to accounting
in subsequent years.30 Some
authors,
like nierman,31 recognized
the
significance
of these new measurement
concepts at an early stage, but, in contrast
to the mathematical
developments
(mentioned in the preceding subsections),
accountants in general only slowly absorbed
these ideas. Among the sixteen papers on
accounting measurement
presented at the
1965 Seminar on Basic Research in ACCOUliting Measurement32
there was only a single
one33 quoting Stevens' work ane! mentioning explicitly its accounting
applications.
Although today this conceptual apparatus
is more frequently referred to by accountants,3! I have the impression
that miJre
widespread
and profounder
familiarity
'with it could avoid many misunderstandings still occurring in our litcrature.35
In
addition,
it would help synchronize
accounting literature with management
science and the scientific approach in general.
But instead
of following
this modern
trend36 several accoun tants concerned with
measurement
theory clung more to traelitiona I ideas or tried to develop their own
new measurement
concepts for accounting
(i.e., retrospective,
contemporary
and prospective measurement)
.37Suchexpforations,
of course, should not be rejected and may
have a double justification:
On one side it
might be that some measurement
concepts
peculiar to accounting arc needed; on the
other side these newly developed concepts
may \VeIl stimulate
measurement
theory
in general and find broader acceptance,
provided they are in tune with the general
trend.

Bclza~,ioral Research
Accounting,
conceived as a normative
discipline, cannot rely on formal proposi-

tions alone. On the contrary, apart from


some normative statements,
the :;ubstance
of accounting ultimately ought to be made
up of empirical
(positive)
propositions.
Since most of the latter are of behavioral
nature, it falls to behavioral accounting to
formulate
specific
empirical
premises

ed. by Thrall, Coombs and Davis (John \\"iley & Sons,


Inc., 1954); \v. S. Torgerson, Tlteory and J(cJ,ltods of
Scaling (John \Yiley & Sons, Inc., 1958); C. W. Churchm:tn and P. Ratoo"h, elk, M casurclI!wt Deft/li/iolls alld
Theories (John Wiley & Sons, Inc., 19.19); J. Pfanzagl,
Theory of ill easllrcmenl (Physica, 1968).
~. H.. Maltcssich,
"J\lcssung, Vorausberechnung \lnd
lluchhaltung,"
Xci/selIYi!l flier IWIldrls-.cissf1lscltaJlliclte
Fo,schllll.~ (April 1959), Pl'. 179-94.
30 R. :\1atk'Ssich.
"To the Problcm of :\feasuremcnt
~n1l1Statisli~al Estinlatinn
of Errors of Accountin~,))
.\f IlIlagclIlcJlt J /I/cT1l<l/it>J/l/{,VoL 2, '''':0. 2, Pl" Ji-5~, :md
idem., .lcw/tillillg
and .I/lalytical.\leihods,
ChI'\. .;, Pl'.
52-96.
31 Harold
Bierman, ":\fcasu[ement and Accounling,"
'I'm: Accou:->T1:-':GREVIEW (July 1963), pp .. 101-07.
" Reprinted in Resr:arc!t in A ccOllllting J[ ellsllrrmw/,
ed. by R. K. Jaedicke, 'r'. Ijiri and O. Nielson (American
Accounting Association, 1966).
JJ Carl
Devine, "Some Conceptual Problems in Accounting Measuremenls,"
in Ibid., Pl'. 13-27.
,. For exampic, "2.3 1Ieasurement Theory" in Report
of the Commilll'e on Foundations of Acc()untin~ ?\fcasurcment, TilE ~\CCOUNTINGREVIEW-Supplement
to
Vol. 46, 1971, Pl'. 309-12 and 322-23.
:c.The lenl;thy controversy hetween Professor Chambers and myseli in Cost al/d M a"ageme,,/:
Ra~'mond
Chamhers
";\ssct
:\fcasurcment
and
Vahmtion"
(}Iarch-April
1971), Pl'. 30-35, and idem., "?\fe:lsuremcnt and Valuation, Again" IJuly-Aul;ust
1971), Pl'.
12-17; R. :\l auessich, "On the Perennial llisunderst:lndin~ of As,ct l\fcasurement hv ~!eans of 'Present
Valucs"" (:\l:trch-,\pri[
1970), ami idem., "On Furthcr
:\Iisundcrstandings
About '1\lea~uremenl' anti V:tluation; 1\ Rejoinder to Chamhers' Anicle" (\!:Irch-,\pril
!'Ji!), PI'. 36-42 and idem., "Assct ?\fcasurcmcnt and
Valuation-A
tinal Reply to Clumhers" (July-~\ugust
1971), pp. 18-23.
3'l The concept
of "measurement by fiat" (Torgerson)
referring to mC;l!;UfCnlcnt based Oil prag)'w.1ic JrY,botheses
instl'ad of scienlific laws (as in case of "fundamcntal"
and "derived measurement"),
is I)f extrcme importance
for accountant"
hut rarelv fullv underslood. Furthermore lhe concepts of "fon(lamenlal" and "derived measurement" long-estahlisheu in measurement theory arc
often misintcrpreted
by accountants or confused with
primary (direct) and seconuary (indirect) measurcmen t,
respectively. This scems to be the case in the otherwise
excellent "Report of the Committee on Accounting and
Information
Systcms" (1971, pp. 309-10), since accountants hardh- evcr deal with eilher derived or fundamenIal mcasure;nent but with 171(OS1lre",e111by fiat.
r. Y. 1jiri, The F01l1ldaliol1s of ..t cCOlw1i1l1( .If easureme"e Prcntiee-Hall, fnc., 19G7i, and "Report of the
Commillee
on Foulll]ations of :\ccounting :\fcasurement," Tnr: ACCOUNTINGREnEW --SuPl'1cmentto Vol.
46, 1971, especially 1'r. 26--32.

II~; IW~

~,
I

/.;:

.....;;.~t
.t>~~

'~:,,::i:';
';il~~I,
.,'

I
"hJ~;;:;'
n

:'1

f:! !
.

.~Ii
r,I!

'III

t~tyi";'

b.~ilnf.:,,.mrl'

~esi~~
aceounlintii'lM

~1.:e~::E
.~ ..'..

,:

L'i'I

bii\ioritP

~1:i/

f\lID .

,''l
iS~'tties,: dW~~j
books ~U{ L.'
J' 't " .,.:;.
~

wn ten, and at
This border ;l%,,(':
is subject t<J:::.:I.
argument is nol

b~h;l\ioral acco
faulty or sup.
havioral rcsear
accountall ts t
'havioral
scienl
about methode
the lack oi suf
ha vioral re;:c;lI
tants. Further
to draw theu
counting has I
can be no dou
inquire in to tv
and economic
and into the
the users of ae
less unanimit,
question whe
into the bcha
counting 5t:1.t\
into the Leha
A further is
arises from it:
and concerns
"The ideal we
tions of accOl:
of economics
to express thl
terms of PSj

:l\Jattessich;

~lethodologicfd

I'rccondi t ions
j
i

:orn
nce

oral
g to
ljses

!SIlOS,
ds (If
\lfch::and
l7.a;;I,
, und
I:liche
>!n'

::tl

" ,;nd
;, Pl'

-,

.lng,

i;.
:-:,,1(tll,

crican

'~~port
- :\lealllt

to

':ham'rnond
;tlion"

~tsure
ii, Pl'.
111111cr-

-n.~l"nt
,rther
Valua
-:\pril
"l and
~ugust
.t~r
S

:J!J

:ental"
,rlance
"II her.1 nll'a'fy:ue

'with
, cment,
oj

';.t~"""'~~ise

ld

,"<1-">Uf(ell
the
t:asure,lJ Vol.

which ultimately
shoulll fit into a more
general framework.
'
The emergence of this "behavioral
accounting"
indeed is one 'of the'main
features of our transition
period. Although
Argyris
undertook
highly
original
behavioral studies in budgeting38 as long as
twenty years ago, it prob~bly was Stedry's
award-winning
dissertation3D
which gave
the signal to the new trend of looking at
accounting
from a behavioral
point of
view. Especially
since the middle of the
sixties, dozens of accounting
papers and
books with behavioral
implications
were
wrilten40 and anthologies were compiIcd ..1l
This border area still is in its infancy and
is subject to many controversies. The chief
argument is not so much bunched
against
behavioral accounting as such, but against
faulty
or superficial
applic;ltion
of be
havioral research within this area.4~ Nonaccountants
trained
thoroughly
as behavioral
scientists
occasionally
complain
about methodological
deficiency as well as
the lack of sufficient generality of the behavioral research carried out by accoulltants. Furthermorc
the question of where
to draw the boundaries
of behavioral
accounting has been left unans'wered. There
can be no doubt that this subarea should
inquire into the sociological, psychological
and economic foundations
of accounting,
and into the related bekwioral
traits of
the users of accounting iriformation.<l3 But
less unanimity
exists with regard to the
question
whether
it also should inquire
into the behavior of the, producers of accounting statements
anci! systems or even
into the behavior of accounting
II/corisls .
A further issue of behavioral accounting
arises from its interdisciplinary
character,
and concerns
the problem of reduction:
"The ideal would be to express the foundations of accounting measurement
in terms
of economics and administrative
behavior,
to express the foundations
of the latter in
terms of psychology
and sociology, the

foundations
of these again in terms of
biology, then in terms of chemistry, and
finally in terms of physics."-14
But at present, we are far away from
such an ideal, and there are dangers in
present attempts
to reduce accounting to
morc "eknwntary"
foundations.
As de-,
sirable as this reduction and integration
would he, it cannot be donc unless: the'
pertinent
"laws" of the more basic disciplines are perfectly understood and as long
as the causal connections
bctween
accounting
phenomena
and the "laws" of
these basic disciplincs havc been fully disclosed. As long as this is not the case, accountants
must do what the chemists did
before the periodic table could be explained
in terms of subatomic particles; they must
construct
self-contained
theories on the
basis of what is known to them. Then
slowly, step by step (with the progress of
their own discipline as well as that of the
more basic disciplines in which accounting
is embedded)
they will be able to extend
their theory by reducing it to explanations
in more basic terms. I do not think that
2. C. Argyris, The Impact of Budgets on People (Controllership Foundation,
1952).
2' Andrew C. Stedry, Blld~d COIl/rol and Cost BeITorinr (Prentice-Hall,
Inc., 1960).
We may refer to a comprehensive pertinent bibl
i ography recently arranged by E. H. Caplan, J. C.
llirnberg, T. J. Burns, J. J. Cramer, T. R. Dyckman,
W. N. Gentry and R. J. Swieringa for the "Report of
the Committee
on Behavioral Science Content of the
Accounting Curriculum,"
TUE ACCOUNTINGREvIEWSupplement to Vol. 46,1971, especially 1'1'.260-84. Thus
further detailed citations seem to he superfluous.
:
William J. Bruns and Don. T. DeCoster, cds., Ac
counting and lis Behavioral Implications (~IcGraw-Hill
Co., 1969); John W. Buckley, cd., Contemporary AccOlm/ing
and Its EllvirolllllCltt (Dickenson
Publishing
Co., Inc., 19(9), Section 7, "Accounting and Behavioral
Science," I'p. 399-445.
<Z Don
DeCoster and K. V. Ramanathan,
"The Influence of Indi\'idual
DilTerences, Task Attributes, and
Control System Characteristics
on Goal Setting and
Goal Accomplishment
Behavior,"
Seattle, 1971 (Mimellgraphed),
also raise this kind of criticism in the first
part of their paper.
<J See Michael
SchilT and A. Y. Lewin, "The Impact
of People on Budgets,"
TilE ACCOUNTING REVIEW
(April 1970), Pl'. 259--68.
'
Report of the Committee
on Foundations
of Accounting Measurement"
(1971), p. 48.

.. ,. --=,

;~11.b

!!

;~T;r:
',' ., _ If

~latt~~.:~=

this could he done by accountants conjecturing their own sociological, psychological or economic postubtes on a wholesale basis. The foundations of the more
basic disciplines must be worked out in
earnest search by their own experts. Only
then may accountants borrow them and
try to add their own modest contributions
to these parental disciplines. "Starting
with grandiose synthetic Yie"'s instead of
working in a piecemeal analytical way, is
characteristically nonscicn titie .".15
In concluding this section, I would like
to ~mphasize that the adaptation of scientific methods to accounting, so far, pre--lorninantly sterns from "mathematical
accounting." But "management scientists
and systems analysts during the last dec-"
ade have gained two important insight{:,
(1) that epistemological research is indi~ll
per{sable for probing the fundamentdl
problems of the management
sciencd~
and (2) that model building, especially tl~e
construction of systems, may impart ne~v
vistas to epistemology ... also, accoU11tants must determine which accounting
systems should be accepted for a specific
situation and which are not applicable.
These central questions concerning the
theory and practice of accounting ultima tel)' belong to the science of knowledge."4s Thus, in their fascination with
mathematical techniques, accountants seem
to have neglected important philosophic
.aspects. Yet, there is growing awareness
that this situation has to he remedied, and
recent A.-\:\-research reportsl~ and articles
indicate that accountants are perceiving
the most critical spots within those epistemological issues.4S The remaining part of
this paper shall outline the major problems
to be soh'cd in this area.

f1;)

SPECIFICATION

OF OBJECTIVES

~e
need for different accounting models
could not remain unperceived in tradi-

tional accounting. The early distinction


between financial and cost accounting, and
the later distinction between tax accounting and accounting satisfying various commercial laws, are the first modest steps in
the direction of that kind of specialization'
\~~i~~u}linE!!:h:_SOUldresult in a huge.
taxonomic structure. A further, but mucfl
~re 'blurredevlclellce
lies in the weary
controversies auout tile correct accounting
method as to valuation basis, realization
criterion, depreciation and allocation methods, classification schemes, and so forth.
It is "blurred," because rarely is the question asked "which instrumental hypotheses or rules are appropriate for a narrowly
defined purpose and which for'another one,
and so on?" Accountants rather behave as
though there existed only one overall purpose of accounting; Thus the inquiry was
directed toward the one and only correct
set of rules; something bound to lead to'
numerous futile discussions and misunderstandings. However, the fact that such
controversies are still ~ngoing49 may indicate that the root of the problem lies much
deeper than in mere misunderstandings or
a onesidedness of this scholar or that. The
heart of the problem might rest in tile dijjj<>llulll;e, Sriwlific Research I, (l9()7), 1'. 30 .
"Report of the Committee on Foundations of Accounting Measurement" (l97l), p. -lO.
" "Report of the Committee on Accounting Theory
Construction"
(1971).
"Until recently, accollntill~ theory laq:;ely consisted of a set of competing a priori arguments about
the relative merits of alternative accountinl; n:ea5ure
ments. Currently there is a growing body of research
that is attemptin~ to estahlish an empirical tradition in
accounting theory development and verification. A full
understanding of such research re'luires that the purpose of such research be viewed within the context of
the history of the scientific method." "Report for the
Committee on Accounting Theory Construction," p. 77.
See furthermore \Yilliams and Gril1in. "On the Nature
of Empirical Verii,cltion in Accounting," PI'. H3-78;
~tallcssich, "Some Thoughts on the Epistcmology of
Accountinl;," I'roacdilll',s, Sccolld f IlltnralioII 'Il COllfcrt,lCe 0'/ ..!c(nllllliug Fell/calioll (London: 1970), pp. 4655; and Devine, Essays i" Accoltntirzg, Volume 3
(mimeogr., Tallahassee, 1971).
R. Mattessich, "The Market Value Method According to Sterling," A baclls (Australia), (December
1971), pp. 176-93.

culty...
to

. ';~';~I~jfP
fC1i!t )....J;ii:

_:Ii

poses, ar:d :" 11lIi~;"


lrypolJ:tsrs.
::::
This t3.:'',,: ~
experts det:m il bll

oy clinging to thc:i
purposc accountin
these pessimists"ai
of acadenlic acc6ui
t,o predominantly
\\rho assume a '111
view, bclicvc that
culty of a: problerr
a challenge than
able number of
seem to :be unc{
perhaps not reee
involved.) If op
because account~
stuck in the dilel
tives and detern
for a spe~ific pur
lenge with. mal
systems experts
The most pr<
problem still see
a limited numb"
poses (tax acco
istic accounting
control, long-n
decision-making
rent operations i
from which a
purposes may I
concepts and ~
quircd for diff
purposes respec
previously men
hensive taxono
concepts. Henc
a taxonomy ,vo
to a Izierurclzy 0
The difliculty
variety of concept
aware that throu
ferent method; 01
terpn:tatiolts of t
I

culty to formulate
~o

specific ~eell-defi:llf.1L/2.1J.J:.=.may comc into being. If olle adJ~ to this three differcnt valuation methods, thcn I'.'e attain, preII specific set of
cisely spcaking, as much as !line different interIt)' 20U!f~,L
pretations of income and of capital. But we know
-This
task seems so diiTlcult that many
that thcre exist more than three depreciation and
experts deem it hopeless, and seek refuge
valuation methoJs, and also that, many more
by clinging to the ill-defined uni- or multifactors beyonJ depreciation anu valuation methods atTect thc conccpts of income anu capital.
purpose accounting system of the pasL If
The hYI"'I hest's of rt'alizat iOIl, classification,
these pessimists arc right, then the future
data-input, duratiou and relevance arc just as
of academic accounting would l,e restricted
indi~l",nsahle for a precise interpretation of thcse
to predominantly
legalistic is,;ues. Thost:
two cuncepts as are ueprcciatioll (i.c., allocation)
who assume a more optimistic
point of
anu valuation_ From all this might result a huge
maze of subconcepts or interpretations, the use
view, believe that the high degree of rliftloi
\,'hidl Ill;ty lIot illlllll"liiatelyI'e obviuus.
culty of a problem shoulJ rather constitult:
But it has to bc pointcd out that these concepts
a challenge than a deterren t. (.-\ considerdo not come into being throu;;h the suggcstion to
able number
of accountants.
however,
],uild a hil'rarchy of concepts_ These concepts
seem to be uncommitted
in this regard,
ha ve beclI exislill,~fur consiuerahle time in the ac
counting literature, sincc one actually operatcs
perhaps not recognizing the crucial issue
with many of those cumbinatiuns. Thus ou, suginvolved.) If optimism
is justified,
it is
gestion docs not so milch aim toward the creation
because accountants
;lre not the only om's
llr in\Tn t i, In of these cltlll'Olll-q>lshu t towanl
stuck in the dilemma of idl'ntifying ohjt;,-making them aware and sy,;tematizing thcm.'o
tives and determining
the optimal model
for a specilic purpose; they share this chalHut the problem of specifying objectives
lenge \vith management
scientists
and
is not only tied to that of a conceptual
systems experts in general.
hierarchy, it is no less closely related to the
The most promising approach
to this
ability of testing whether a purpose has
problem still seems to be the acceptance
of
been fulJilled by a specilic (information)
a limited number of typical standard pursystem.]
n this an:a accountants
have
poses (tax accounting,
commercial-legalgrea~ly benefited
fro~ inform~t.ion
eco(I
nOllllCS and mad\.: llll:lr own onglllal conI'stl'C accountilllTb' short run l)l'lnllill
'b
,111
I 'I
.
.'
.
control,
long-run
planning,
inyeslment
n )UllOn-a
further reason for optimIsm.
T'

\f1-

Jut
"C

cch
, III

_,ull
~Ir-

'-of
.he

decision-making,
decision making in current operations including personnel policy)
from which a larger number
of sul>.
,.,
.-purposes may I)e derIvnl.
:-'Int"e
dllkrent
concepts alld subconcepts
would be required for different
purposes
and subpurposes respectively, we come back to the
previously mentioned need for a comprehensive taxonomic
syslt:m of accounting
concepts. 1knce Ihe construction
Df such
a taxonomy w0111d ha \'e t D Iw do~-wlv Iit,d
to a hierarchy of objccli,'cs.
The difliculty of such \lndertakin~ lies in the
variety of concepts and siluali'llls. \\""l>llght tll he
aware that through the al'plicllion of llIree dif.
ferent methods of depreciation three dilTerent interpretations of thc incomc and capibl cllncepls

(Lj)TESTI~G J\rANAGE:\IE~T

P
[C'

11 ,)(,<IT(

hFOR~IATION

SVSTE;\IS

['"

of

Tlllnj'!n

'j

Of

'j,~

..,
I csll/If'

Systematic
testing
procedures
arc
crucial clement in any scientificapproach.ol

'" TranslatcJ from:'fattcssich, {)ieu'isscllsclwjtlicllt:n


Grulldl"gCll des Rccll1lull~"cr:sms (Ducsscldorf:llertelsmann l-nin'r,itactsvcr!ag, 1970), p. 46_

C,l\\"illialllsalld (;rillin, "(In the :\alnrc of Empirical

\crilic.atioll
ill :\c("()lIntitl~,"
p. 1-13,;ts~("rt that l' ... only
Inil1illlal
alh:nliol1 lIa~ becOl1 ~ivcn to fpH::-;tioll~ of t'r,ifi~
(:rltiOll.
or fluor'! 'i.'ll/idalio1t ....
But no scriou$ atH

tempt has I.n"n-madeto identifythat whichcOllstitutes


\-nilieatiulI_1l"wever,the rc;uJcr'saliClItionshould be
drawn tu the hct thall de\,()(cuthc clltire Chapter i of
..1,coIITIling ,,,"I ..lTlalyliCilI
.!felhods (196-1) to empirical
hYI'"thc,;,-s:llId llwir refutation. The fulllllvinl;quolc
SIIIJllllari/.t:"s

till'

pertinent

CtJIlc1u5ilJu:

... __ehu;c", Ilctwccn :lniun hYI'0thesegwill be


lIlaJe ill such a way that th03ChY!,othest:s--!towlow

The cogmtlve sciences must test hypotheses and theories bv means of verification
and refutation; whereas the applied sciences ought to test the efticiency, relevance, reliability or other properties of a
normative theory, or a system, or a machine by a variety of means. }'Jechanical
devices, like motor cars, are tested in many
,'/aysand with respect of many properties.
Some or all of these properties (price, relative gasoline consumption and other econpmic aspects, motor strength, acceleration speed, sturdiness and further safHv
features), or their combinations, are tlle~
compared from model to model, and co'mPilred with an "ideal pattern" determined
by the intended usage. Only then a correct
c~oic,e for the appropriate purpose cani be
nlade. Most readers are quite familiar ,,;ith
this kind of "testing" from their experience:
in buying various gadgets. Although hccounting models or information systdms
ate more than mechanical devices, th'ere
elm be little doubt that they too belong to
the realm of applied science, and thus are
subject to similar principles of testing.
Obviously, this testing can span a wide
spectrum of rigor, from the most informal
estimation of certain properties to highly
sophisticated measures; hut the marc complex an information system is, the less
satisfactory "'ill an informal testing procedure be. Thus, under consideration of expensive computerized
information svstems, considerable thought will have to' be
given to the procedures of testing and,
above all, to the principles underlying
them. In the pure or cognitive sciences
these "principles" are thoroughly examined in epistemology (the science of knowledge) . We too are concerned with a kind of
knowing, namely with the question "IIow
do we know when a system is satisfactory
or optimal?" Therefore, there is no reason
why this branch of philosophy cannot aid
us in our undertaking. It is true that so far
no epistemology of the applied sciences

exists, but systems analysis, management


science and recently accounting are posing
fundamental questions that most appropriately would fall into the epistemology
of applied sciences (if this expression
sounds too highbrow for accountants, the
term "instrumental reasoning" may well
be substituted for it). Some of these fundamental questions are not new at all and can
be encountered with Kant and many other
philosophers.52
Other of these questions have a more
novd aspect as the following example shall
demonstrate:
Information is knowledge
and thus there exist close ties between information theory and the science of knowledge. Yet, these ties have hardly been explored. Economists worry little about the
criteria of knowledge creation, and epistemologists care even less about the value
and cost of procuring knowledge or information. The latter issue has recently been
taken up by economists as well as accountants, and if both groups would acquire
some methodological background, then thc
perennially 1lcglected eco1lomic aspect of epistcmology could become a fascinating interdisciplinary meeting place.
Thc Faluc of lItformation
The testin~ of an information system has
many prercquistics; among these the most
their de~ree of reliability might he-will be considered acceptable which, in the ('tce of uncertainty and
economic constraints,
are at least slightly more
I,rolllising or lIlore reaslln:d,le than their alternating.
~rhus we 11fOIH)SC to Cxplol-llhc proccs:; of rnv:tlida.lion
a5 a criterion for di:-;tin14l1ishin~ hetween pra~lnatic
hypotheses (or action hypotheses) and scientilic hypotheses:
"I. A scientific hypothesis is invalidated by instances
(acceptable 10 the experts) which testify reliably to
the falsity of this hypothesis.
"2. A pragm;ttic hypothesis is invalid;tted (rejected)
by demonstrating
(or believing) that, ill the lo/zg rtm
or on the average, the actions hasell on it yield results
Lhat ;trc less satisfactory than the results of actions
hased Oil allother, avail:tble or procllral,le hypothesis." (1'. 235)
C. W. Churchman's
discussion of Kant's imperatives of skill and the asscrtorial imperative, in Prediction and Optimal Deci"io,z (Prentice-Hall,
Inc., 1961),

pp.31-36.

au
fOnn:i~'

m"

tern.1
the~l!
tion~'
wheth&t"""
...,.,.~
~or wheth;
~e

'.

sam~'inf:
'''nlll~. H'1i

1nmany:

;t'termin!
he co~t f"'"
'.W

'il

:flusiv~:~c.i~.
.'
":lr ..:.~.fi.'.t~.
tion. Recent "st,l'.
nomics h~ve~~
formulation of ie~
lern, tO~l~hin~i'~
and . mana eDi,a,
.'
;:, .. g ~
analys~~.i'fhu~ac
to be aware ohi
:.'" kind odesearch'
tion to put its
text. Thhm~
one sidc,'to relab
more traditional
and on 'tile other
sophie roots of
Information
tion as a resour
and related fea
natural exteasic
theory which, an
to improve prior
additional info
former into post
ian approach).
much is this
worth?" alread)
theoretical issu
, elaboration. Fel
an exciting con
operational info
accounting situ[
, specific model s'
plifying assuror
signed for imme
practice. N evert
the best exampl
l~

r;

.1
I
r
i

,I

i
1

crucial is an awareness of the value of information procured by the pertinent system. Without some notion of the hCI1l't":t,
the cost and the net value of the infUl:mation created, it is,neither possible to test
whether this information is worth creating,
nor whether another system could create
the same'information more efTlciently. As'
in many other cases of systems'; operation, ,
the cost of doing !iO is relatively 'easy to de-,
termine-at
least compared to the often'
elusive benefit of gross value of information. Recent studi~s in information economics have contributed toward a rigorolls
formulation of several aspects of this problem, touching the very core of iccountillg
and management
information systems
analysis. Thus accountants not only have
to be aware of the present stages of this
kind of research, but should be in a position to put its results into a broader context. That means they should be able, 011
one side, to relate these new insights to the
more traditional aspects of their discipline,
and on the other, to understand,the philosophic roots of these insights.
Information economics regards information as a resource having a value, a cost
and related features.S3 This branch is a
natural extension of statistical decision
theory which, among other things, attempts
to improve prior probabilities by means of
additional information,
converting the
former into posterior probabilities (Bayesian approach). Thus the question "lIow
much is this (additional)
information
worth?" already looms in m:my decision
theoretical issues, even without further
elaboration. Feltham and ])emski;'" made
an exciting contribution by constructing
operational information models adapted to
accounting situations. Even the applied or
specific model still has to make many simplifying assumptions and is harcily dcsigned for immediate application in actual
practice. Nevertheless it constitutes one of
the best examples of serious accounting re-

search performed in recent years (indeed,


the paper received the l\ICPl\-Award for
JIJ7()
and ought to 1.)(; taken as a prototype
for the kind of research (at least its analytical part) desirable during the "transition
period."). It demonstrates that fundamental research in accounting must start from
simplified situations which then have to be
refined slowly step by step; it also shows
that the decisive matter about fundamental accounting research is not an immediate practical application, but conceptual
and methodological clarification to attain
new insights into the complex relations of
information creation allu evaluation.
But there exists another recent publication (incidentally also an award winning
article: one of the winners of the American
Accounting Association l\Ianuscript Contests for 1971) which supplements the
above mentioned research in an important
way. Mock's papers; draws attention to
the fact that the concept of information
value used in information economics is
only one among several concepts or interpretations, and that it is paramount for
accountants to distinguish clearly between
the three types: the economic value of information, the model value of information
and the feedback value of information.
J\lock's cbssification of the value of information concept may not be the ultimate
solution, but it seems plausible that the
present economic information value concept alone provides too narrow a basis for
management information systems; surrogatc valuation methods might havc to be
incorporated.
OJ Henry
Theil, Fcollomics al/ll Information
Theory
(Rand McNally & Co" 1967), and James C. Emery,
Orgllllbltivllal
Plamlill,~ alld COlltrol Systems (\fac?-lillan Co., 19(9), PI'. 66-107.
"G. A. Feltham and J. S. Demski, "The Use of
~todels in Information
Evaluation,"
THE ACCOUNTING
REVIEW (October
1970), PI'. 623-10. See also idem.,
"Forec:lst
Evalu:ltion,"
THE ACCOUXTIXG REVIEW
Uuly 1972), PI'. 533 .
Theooore J. ~lock. "Concepts of Information Value
and Accountin","
THE ACCOUNTIXGREVIEW (OCtober
1971), I'p. 765-78.

Finally one may take into consideration


the fact that information theory and information economics, thus far, have been
treated on a predominantly
stochastic
basis.56 Thus the question arises whether
additional interpretations of the "value of
information"
also require probabilistic
models or whether some further a~pects
can be treated deterministically.57
I
Apart from the interpretation
issue a
further important problem emerges::What
other criteria besides the "value o( information" are required to test an accounting
system? It seems that there exist. many
more criteria indeed; for example the degrees of relevance, of efticiency, of ei1ectiveness, of accuracy, of reliability, of timeliness and so on. These could be taken into
consideration as additional qualitative
characteristics,5R or, as informatiOlt economists would suggest, one could incorporate them into the value of information
concept. In other words one would assign
a higher value to a system with higher degrees of relevance, eflicicncy and accuracy.
To what extent and in which way these
criteria can be incorporalcll is still a matter
to be explored. One possibility might be
the assignment of multi-dimensional
utilities to the information model or of (listinguishing a greater number of information signals than otlwrwisc. \\'ith n~~ard to
"response time" (degree of timeliness)""
the problem could be solved by constructing dynamic information models, i.e.,
models taking into consideration the time
dimension and thus the interdepcndencies
between information and lH.:nclits.loC<l\cd
in dilTerent time segnll.:nts."o
.
Among other vehicles for testing systems and operations research models, scnsitivity analysis is mentioned most frequently. Indeed, it is a necessary but,
often, not a sufficient lool for testing a system. As far as informa tion systems are concerned we may refer to the AAA Report on
Accounting and Information Systems:

Observe that sensitivity analysis is asking cssentially the same type of questions about information that were considcred informatiun economics.
However, sensitivity analysis is a more ad hoc
way of asking these questions. The basic rcason
for the more ad hoc approach is that mathematical models usually assume that the parameter
predictions are deterministic whell, in fact, they
are uncertain. Recognition of this uncertainty incrcases the complexity of the models and the
resultinR increase in the cost of solution may not
bew~ted.GI
~

THE GENERAL THEORY

i\:ccounting research during the last


fifteen ycars not only greatly maturedlJut
also spread into many directions. The centrifugal force at work in our discipline during the transition period is well manifested
in the great variety of modern accounting
topics. This force and the dynamics behincl
it mi~ht prove for a.cc()untin~ either hi~hly
heneficia lor destruct ive, depelldin~ on how
accountants are able to harness it. If the
many fugitive parts and pieces of our discipline can be held together and integrated,
accounting as an academic discipline will
survive, if not it mi~ht dissolve and be
ahsorlwd by nl'i~hh()ring fIelds. The present state of accounting research rcsl:mblcs
a jigsaw puzzle where some areas slowly
grow into meaningful configurations but
"The information

theory approach to the 'luantifii$ ha~cd upon the prcnli:;l': that, for
:tlly
plohkul. there arc :t cerlain IHuuber of possil,le
;tn,;wers to which plohabitilics may be allachcd. When
inforlll:ltion ahoutthe problem is receiv<;d, the original
probabilities
un<!erl;o transformation,"
Baruch Lev,
.recoll"till!: alld hzformation
Theory (AAA: Studies in
'\ccounlin::; Research, 1969), p. 1.
.1 The solution proposed bv Theil as well as Lev, to
introduce "artificially" probabilities by assi~ninp; them
to individual asscl items on the arbitrary basis of the
\"alu<;ralio of lhe pertinent assel which it has in proportiou to lhc balance shecltot:tl, is the attelllpt o[ a compromise .
Emery, pp. 98-107.
.0 By up dating or procuring information quickly the
value of this information can be greatly enhanced in
somc but not in all cases. See "Advances in Fast Response Systems," [<;f>P-AlIalyzer (Fehruary 1'>67).
Gerald A. Fcltham, A Theorclieal
Fratn~J.'OTk for
Evalllali"g CI"!IIges ill A ee01t1lli"R I llformatioll for J[ anagcrial Decisions. Mimeographed Dissertation (lJerkeley 1967), especially Chpts. V-VIII.
.1 TIlE ACCOUNTING REVIEW-Supplement
to Vol.
46 (1971), p. 307.
Cltttlll

or infotlnalinn

without yielding the entire picture. Indeed,


the individual fragments
seem to spread
outwards and not towards a common center.
The need for a deliberate elTort of integration in our discipline is by no means
of recent vintage,
but it becomes ever
much more urgent ill a time of opulent
growth and specialization.62
The many cndeavors to formulate
the basic assumptions of accounting
fusing them into a
postulational
system goes back 1ifty years
to Paton's first attempt.63
The search for
accounting postulates since the second half
of the tifties hy Aukrust, Chambers, Ijiri,
Moonitz and others, gives evidence for the
acu te a wan:ness of these neelb.,t :\s long
as we do not possess an O\'erall accounting
theory, a good deal of the advantages
of
mental economy and generalized thinking
will be lost. Where no general theory is
available, speciJic theories have to lie dLvised anew, over and over again, Lvery
time a specific accounting
system is employed in a particular si tuation. I n such a
situation generaliz~ttion,
the quintessence
of the scientitic approach is ddeall-d, :tnd
all other endea vurs, like beller COl\ccpt u:dization and interpretatioll
and rigorous
employmen t of scient itic methodo!pgy,
lose a great deal uf their ultimate rtIiSIJll
d'Ctre. I also believe that this concern for
integration
should not be restricted
to a
few specialists.
En'!"y an:oullting
rl'searcher should have at least a rough yision
of the overall framework
into which his
specific research \\'ill han~ to fit SOOIH:ror
later.
To some extent not only postulation
of
business accounting
(or national accounting in Aukrusl's case), but integration 011 a
broader basis has been undertaken
in accounting over the span of the last lifleen
years, The first proposal to integrate
all
areas of micro- as well as macro-accounting
by means of a postulational
system was
launched in 1957 and further elaborated in

196-! and 1970.";' In most recent times a full


in legra tion of micro- and macro-economic
input-output
analysis,
with special emphasis on the accounting aspects, has been
launched
by Butterworth
ane! Sigloch.6&
Even more recently this generalized inputoutput approach has been fused in a most
inlL:resting way to the theory- of information economics.6'
In such cases of overall
integration
the most crucial qUl:stio!1 concerns the empirical premises. Can a truly
general theory of accounting
be successfully constructed
on the comparatively
narrow empirical basis of decision theory
and information economics, or do we need
a larger number of factual premises!' Those
a[(lrlllillC!; the first part of lhe question
might claim that behavioral accountants
ha VI: not yet supplied any empirical hypotllL'SL'Ssuitahle as foundation stones of a
general accounting
tlwory. While those
""\\hile
much research h:<s hecn focused on the
anal,,is of minutc sc"mcnts of knmdedc;c, thcre has
laTH illcrl':lsin~ inlcre~t in ltcn:l()pin.:~ br:,.:<:r franles of
reference for s,nthesizin~ the results <,f such rcscarch.
The allention has heen i"Kusel! morc 'l'1d more on o,cra!ls,stems as Ir:<mes of refnellce for :1:1:<',tical wnrk in
\';lri;t\l~ :In:l~.1' R. :\. jollllSlllt,
F. L. K~lSl and J. E.
l~t):;('ll:.\\t'i~.
"Sy~~Il"fI1S
'I"IH'III.\' :lll,i
\l;UI:q:clIlcnt,H
'\11111,1~t:m<'ll1
Scitllre (January
11)()I), 1'. 3(,S.
1\:1 \v. A.Patoll)
...lccollHliug
Thror}' (l-:'ol1ald Press Co."

1922) Ch!>\. 20.


O! John'
W. Httcklcy, Panl Kircher "nd Russcll L.
.\lathcws, ~lcthoJolo;;y in .:\t:C()Ullttl1~ Thcnry/' 'rllE
'\CCOI;:-;TI:-;G
REVIEW !.\pril
1')6S), PI'. 2T+-83. Thcse
authors n"t on!v criticize the deficient delinitional, concc-ptu:Il, "lid n',elhod"logictl
'Hlrk (,f traditional
acCOlll\tin~

rcsc;lrch.

but abll

put great

cmph:tsi::;;

no the

IllTd for an over :<!Ithn>lTticd structure. For a further


plea in fa,or of thc axiomatic method sce Georgio Pdlicelli, "The Axiomatic ;\Ietho(l in Bu,oiness Economics:
A First ,\ppro:lch," Abacus, (December 1969), pp. 119-

31.

" R. }'httessich,
"Towards a General and Axiomatic
Foundation of Accountancy,"
..1ccollllting Reserclt (October IOS,), pp. 328-55; idem., .lcrotin/illl; t1Jld Alltllytical
JJ d/wds (I~ichard D. Irwin, Inc., 1%1); idem., f)i~ wissOlsclwftliclim
(;It"IIIl<l~m des /{cclinIlJlgso,esClls
(Berte!SInann liIli\Trsital"tsverl;l1.~,
1<)70).
'" John E. BttUerworth and Berndt A. Sigloch, "A
Generalized
}'Iulli-Stagc
Input-Output
"'lode! and
Some Dcrived Equivalent Syslcms," TilE ACCOUl\TI:-;"G
],EVIE\V (October ['Ji!),
PI'. 700--16.
" See John E. Butterworth,
"Thc ,\ccounting System
as an Information
Function," Working Papcr Ko. 90,
Faculty of Commcrcc and Busincss Administration,
(Univ. of British Columbia, 1(71), to be published in
lOllT/wl of Acrowrti>:g Rescarch (1972).

ventures, which arc lIothing more than


such experiments, tried to solve this probafflfming the second part point out that
lem of combining uniformity \vith variety
the empirical hypotheses of decision theory
by ofkrin~ a framework consisting of two
and information economics (mainly the
kinds of prcll1ises: (I) basic assumptions
assumption::; lJorrowed from mo,krn utilamI (2) spccilic or auxiliary assumptions
ity theory) serve only ;1limitctl number of
(abo called speci lie hypotheses) .6~ The
accounting needs, and arc irrelcvant or
basic assumptions manifest the general
even contradictory
to many situatiuns
characteristics of all accounting systems
which a general theory of accounting is
and provide place-holders for the specific
suppos'cd to mect. I consider both answers
assumptions. The laHer l'nablc adaptation
unsatisfactory because they ,10 nol lea,l to
t.o a variety of particular purposes through
;1unilication but ultimately end in a fragmentation of accounting theory. A:::.pointc(1 the choice of exchangalJlc alternatives
(these specilic hypotheses will be shortly
out previously, 1 am convinced that inputdiscussed in the following subsection).
output economics and information ecoContrary to some erroneous opinion, most
nomics made and will continue to make
of the basic assumptions were formulated
most valuable contributions to modern
as empirical propositions which are, unlike
accounting theory, but 1 believe that actautologies, ,not 'true by virtue oftheii
counting and' management! information
:logical ;structure. 'Most of these basic as~
theory <in genernl is confron,-ed' with ad: sumptions are existential propositions and
ministrative tasks for which the behavioral
basis of 'economIcs alone is in~dequate. Ob- I possess refutability, the best witness for
viously,' the easiest way out of .this dilemma
I empirical content.
!
Although the literature of managcmetit
would: be to develop various accounting
inforniatio~ systems persistently points
theories for .different purposeS, indepenthe need for specifying the purpose of ,a
dent of each other. This I \~'ould consider
management information system, tradian intellectual defeat for our, entire distional accounting often failed to do so.
cipline, especially at a time' of progressive
Therefore, Die wissenscltaftlichen
Grundintegration of other areas. Sometimes delag en introduced a neW basic assumption
feat is inevitable-perhaps
our notion of
(only implicitly present in Accounting and
accounting as a single discipline is basically
11nalytical Methods) which requires the
misconceived-but
a defeat without seriarticulation of the pertinent information
ously trying to counter it (that is without
purpose through an additional specific
trying to integrate all of accounting),
a.ssumption (to be chosen from the many
ought to be unacceptable for the aG1demic
community of accountants. Thus 1 conalternative ohjectives).
sider the fonowing two questions vital for
the future of our discipline: (1) Is a.unified
I Illerpretation through lllslrumelttal
H yor general theory of accounting possible or
potheses
not? and (2) If it is, then to which degree
Whereas the basic assumptions ought to
and in which ways is integration feasible?
constitute a common frame of accounting
A solution to this prohlem would require
much experimenting. That means, a series
I was especi;tlly pleased to find my approach of disof attempts should be made to construct
tinl;uishing between these two kinds of premises conand test alternative, comprehensive, thefirmed, as methodologically
sound, in one of the leading
works on the philosophY of science: Bunge, pp. 402-403.
oretical structures amenable to manageSee also "Report of the Committee on Foundation
of
ment information systems, accounting sysAccounting Measurement"
(1971), pp. 43-44.
tems and subsystems of them. 11y own
.

.t

at

-r ~

1\1 t tcssich:1\1

than
prob:lriety

,I two
,', ns
\ltions

. The
"lll.'fal

,;,ems
'ecifjc
f\'dion

r;mgh
atives
'It1rtly
lion).
, most
,iated
~e
\lleir
it' as" :IIlJ

,:; for
.:nt
l'll

ts at

~ of a
!

radi-

if)

so.

:rulld-

:ption
'g 111ld

.; the
;;\tion
I:.:ciJic
many

c;ht to
'lIlting
dis:~~s
con:""ding
!!2--I03.
;lion of

"O(

'

I
clhodologiCal

Precondi

lions

sy~tems in general, the; specific assumptions may serve to give iinterpretation


to
the overall theory. In tljs connection we
may refer the reader to !the beginning of
this paper where the neglect of and need
fori'interpretation"
in accounting was discussed.69 From the very ()utsd of Illy proposal I pointed at illustrations
of specific
hypotheses
and repeatedly
stressed their
, exigency by such phrases as:
While the preceding
assumptions
(containing
a priori as wcll as empirical
n'otions) arc of truly
general
nature,
a series of secondary,
cmpirical
assumptions
(in lhc folluwing
called hypotheses)
are required.'"
But olle of the important
tasks of accounting
theory is the formulation
of various
alternativc
sets of hypotheses
required
for specific purposes.
Yet, before this task can be explored
(see Chapters 7,8 and 9) it was indispensable
to acquin.: a
clear notion about
the foundation
on which our

<l ~ciplinc resls:(


To disregard these specific empirical hypotheses means to misunderstand
the very
essence of the theory outlined in Accounting and AnalyticalllfetllOds.
Yet Williams
and Griffin seem to limit their glance only
to its mathematical
aspect and say" ...
the mathematical
formulations
of the accounting processes by l\lattessich
(a footnote refers to A ccolinting ilnd A /lalytieal
A[ etllOds)-are exa mpks of thcorct ica I expositions which rely for their validation on
the truth criteria of mathel11atics.";~ But
the theoretical structure, presented in A ccollnting and AUlllyticall1fctllOds
as well as
in the revised German version, must be
verified or refuted through empirical means,
above all by testing the specific hypotheses
-something
that cannot be cmphasi;oed
strongly enough.
It may be of interest to learn that some
auxiliary propositions
can be formulated
either as conclusions
(theorems)
or as
auxiliary assumptions
(obviously, only in
the latter case do they belong to the set of
premises). Whether the one or the other
alternative is chosen might be a matter of

taste, bu t definitely depends on the inclu~


sion or exclusion of a basic assumption calling for articulation
of the information purpose as a placeholder. If this assumption is
explicitly stated then some of the auxiliary
propositions
might be derivable as conclusions; if it is not included in the general
part then most of the :tUxiliary propositions become premises (from which further
specific conclusions will follow). This latter
version was used in Accounting and Analytical it! ethods in contrast
to Die wissenscltaftliclten Grl/1uUagen des Rcc1mzmgswesellsi3 which uses a general placeholder
for
speci1ic goal propositions
and therefore
can tains 19 instead of 1S basic assumptions. But the latter version therefore requires a smaller number of auxiliary premises and thus may be deemed more elegant.
Conclusion and Testing of the General Theory
On the basis of the preceding subsections we may now draw some major conclusions and formulate,
what I consider,
the chief message of this paper:
1. The testing (verilication'or
refutation) of a general theory of accounting
must invariably oe tied to the specific empirical propositions
governing
particular
accounting systcms fflr Use in adual practice. l\ccounting
cannot be anything else
!Jut an empirical
(l1ormative)
Hiscipline,
and hcnce th~re cannot exist purely abstract
or nonempirical
accounting
theories; 'what may exist arc uninterpreted
The reader ought to lie aware that interpretation
of a calculus or unintcrpreted
thcory may he achieved
fly either a ,e[',nate
set of rules of interpretations
or by
incorporating
these rules as premises, In the latter case
Carnap speaks of mcaning pastulat",,; see Rudolf Carnal', II/troduction to SW/ll/lties (Ilarvard
University
!'ress. 19-12).
,. Aceount;n!: and Aualyt;Cilt Methods, p.-II.
" Ibid" p, 45.
" Thomas II. Williams and C. Ir. Griffin, "On the
NatUle
of Empirical
Verification
in Accuunting,"
.-l&(/(1/S (I kcemher
1\169), p, 153.
" Chapter 4 of this book demonstrates
in which way
specific or auxiliary propositions can be derived as conclusions from the basic assumptiolls including the goal
assumption.

~--1t
i.

calculi74 of accounting, but they are meaningless unless at least provisions for possible interpretations are made (as pointed
out such provisions were mad~ in Accounting and Analytical M etlzods ani:! more elaboratelyin
Die wissenscJzaftli~lten Grulldlagen desReclwungs7.!.'esens).

I'i :

2; A !particular accounting' system is


tested by trying to determine systematically whether it is the most "~atisfactory"
system "under the circumstan~es," i.e., for
a well-specified purpose. That this involves
many conceptual and methodblogical ditliculties was discussed in the present paper.
Rut the statement holds quite'inclependent
of these difficulties. In my view it is this
testing of particular accounti;lg and other
management information systems which
ought to be the central concern of theoretical as well as practical accoulltant~.
3. A general theory of accounting must
be created in a recursive but noncircular
way: starting from the basis of an operational definition of accounting (sec the
$ecolHl section of this p:lpcr), fairly sharp
!Jut preliminary houndaric~ ()f what constitutes an accounting system should I)c
drawn on the basis of past experience and
existing information needs. Then, all sYstems
obeying the conditions of the oper;tional
definitions ought to he tested. Tf some of
those specitic accounting systems do not
fulfill the specified purpose to the degree
desired, the "why?" must be determined.
If the failure is due to some auxili:try propositions then thesc hypotheses have to he
exchanged or amended, thus the structun:
of the specific accounting system, hut only
of that, changes. If, ho\\'cYer. the failure is
due to some basic assumption, the latter
has to be exchanged or amended; thus the
structure of both the general theory and
the specific system will change. In this way
the general theory is verified or rdull:d
every time a specitic system is put to test.
This testing oj a general accvltllting theory
by u'ay of the empirical -.'eriJic:ation or re!lIta-

tion oj its interpreted systems, seems to me


the only way of conforming to tlze requirements of an empirical disciplille. Thus fac-

tual refutation, even of a general theory,


is conceivable and circularity is avoided.is
Indeed, refutations and consequent revisions of the general theory arc expected,
perhaps frequently in the experimental
stage, and hopefully less frequently in the
mature phase.
4. The methodology thus described is
independent of the boundaries of the general theory. Thus by choosing a narrower
or broader set of basic assumptions, accountants may agree to concern themselves
only with double-classificational models or
''lith management information systems of
wider scope. Most (but certainly not all) of
Illy \,:lsic assumptions will hold for any
Illallagclncllt information system which requires a Ilumerical scale for measuring
values, scales for measuring quantities and
time intervals, economic objects, subjects,
entities, structures amI husiness events,
articubtion of a purpose :lnd interpretations of "value" and "incollle," a c1a:;sification system, a certain (kgree of aggregation
and some kind of periodization. These!
basic assumptions may not be very exCltlili i
r;:;;nl a hehavioral point of ,:ie;:.l..J!.utthey I
7i rc"an-;;;-tcTIcntrncanslO-,(-ou ntc ract the!
Cl'11 (rIf\ig'iif-tcndencles
or modern-7~~nt-f
--c_ ..
jii:r ,lI1(Tr"Cfifca- areas. Furthermore,
the'
_,.--,..".",,,,,,-,.,,---,,-.-<;,.,,,,,~.,...~,

.:_~,~,,,,,,,-.~--~~

"Without interpretative
rulcs, a theory would reduce to an uninterpretcd
calculus which points to
nothin~ heyond itscif. It would hc as cmpirically irreievant as the gallle of chess!' C. J" \fasscy, "Professor
Samuelson on Theory and Rcalism: COlllment:' JlmcriC,II< Icconomic
Raiic:v (Dccember 1'1(5), p. t 159.
Circularity for examplc would hc involved in thc
following situation: (I) The gcneral theory and analytical defmition stipulate the basic assumptions (conditions), (2) the basic assumptions detemline the specific
system, and (3) the gcneral thcory is vcrified hy testing
whcllH'r Ihc specific syslcms fulfill the condilions stipu
lated in the analytical defmition. The al.l.cnlivc read
will have noticed that my s\l~geslion for verification
f
the general theory is radically different from this ci uhr one, uccause the formcr depends vn the fulfillm
lof
u-ell-defined factllal purposes by lite interpreted sy ems,
whereas the latter merely hillges 011 the ouserv' ee of
JclinitiOllal condition;.

articulation of these conditions in a general


tinent results to a theory that provides
way and their subsequent interpretation
testable management information systems
were previously not self-eviJcnt at all. For
for specific o!Jjectives. Academic history
instance, several of these conditions like
offers ample proof that only through cospecialization of purpose, interpretation of
operation and mutual understanding of
value, economic subjects, and degree of
observers ane! experimentors on one side
aggregation were made only implicitly or
and theorists or formalists on the other,
quite unsystematically in traditional accan knowledge be advanceq. in the long
counting.
run. Kepler could not have developed his
5. The formulation of hypotheses for , analytical scheme of the planetary orbits
specific accounting purposes through bewithout the observational results of Tycho
havioral research is (together with analytiBrahe, nor could Einstein have developed
cal research and methodological clarificahis relativity theory without any know Ition) one of the major tasks faced by future
edge of the experiments of Faraday, Galilei,
accountants. Williams and Griffin offer a
the lvlichelson-Morley team and others.
most revealing classified survey of major
Although accounting is no pure science but
empirical research projects carried out durpursues a predominantly
instrumental
ing the second half of the sixties}6 By
task, as an academic discipline it has grown
glancing at this surveyor scanning recent
complex enough to justify a division of reaccounting literature, the reader will notice
search according to inclinations and talthat most empirical accounting research is
ent5. Such a division of lahor is as indisof peripheral nature from the viewpoint of
pensable as is the close cooperation bedesigning accounting systems. That is to
tween those specialists.
say, there exists an insuJlicient number of ~tha--r-t-lc---ta-s~k-'
-O-f-uUl
empirical research projects that deal wi th I/ transi tion;e~iod iie~in the slow bu t serious
such important questions as the following:
endeavor to convert rules of thumb into
(a) How purpose-oriented are users of atwcll-grounded,'7 purpose-oriented or incounting information, and which kilHi M
strul11cntal hypotIH~sls. SUlh a process of
information would they desire or nlTd f(i.r
cOl\vlTsiol\ must not 1)(' I,;\sl'(! Oil the falwhich purpose? (I What is the eried of
Iacious notiol\ that action is the lest of thedifTerent interpretations of value, income
ory, hut rather on the insight that action
and realization on users of accounting, ancl\
can be improved, in the long run, by cogniabove all, which interpretation
matches'
tion of relationships "'ith high degrees of
which purpose? (c) What is the effect of
reliability.
different
allocation
and classification
.. lams and Griffin, "On the ?\ature of Empirical
schemes, different degrees of aggregation,
Verification in Acr:nuntin~/' pp~157 -7g.
rule is gr01wdrd if and only if it is based on a set
different periodizations (interim reports)
of law formulas capable of accountin::; for its efTectiveon users of accounting, and above all,
ness. The rule that commands takin::; ofT the hat when
greeting a laely is groundless in the sense that it is based
which scheme matches which purpose?
on no scientific law hut is conventionally adopted. On
Only if behaviorally oriented accou nthe other hand, the rule that commands greasing cars
periodically is h:lscd on the law that luhric:ltorsdecrease
tants answer these and many other relethe wearing of parts hy friction: this is neither a convenvant questions to a high degree of reliabilt ion nor a rule of thumb like thuse of cookin" or politicking: it is a well-grounded rule. We shall elucidate later
ity will the analytical accoun tan ts and
on the concept of basing a rule on a law." Bunge, Scienmethodologists be able to fuse the pertific Research J[, pp. 132-33.
77 ""